Defining and assessing the spectrum of tuberculosis (TB) disease: application to diagnosis and prognosis

Date
2020-12
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: Although Tuberculosis (TB) has been a topic of investigation for centuries, a satisfactory method of identifying patients with clinically active TB is yet to be discovered. This has two important negative consequences: Patients requiring treatment often remain untreated, and TB treatment is often administered ‘presumptively’ for patients, especially young children, who have symptoms suggestive of TB, even in the absence of laboratory test confirmation. In this thesis, various aspects of the challenge of elucidating the spectrum of TB disease are explored: Key concepts in diagnosis, and defining performance metrics of diagnostics are explored; we consider the challenges of designing a study to evaluate a candidate diagnostic platform; simulated data, structurally aligned to a major ongoing study is analysed to demonstrate options for analysing and interpreting data in the early stages of TB marker evaluation; database schemas for the South African public sector TB treatment programme are explored, in particular relating the different handling of drug-sensitive TB (DS-TB) and drug-resistant TB (DR-TB); a preliminary analysis of DR-TB treatment outcomes is presented. We demonstrate not only that sensitivity and specificity by themselves offer a limited view into the performance of a test, but also that these metrics are not intrinsic to a test, but vary with disease prevalence. More generally, the contextual ‘spectrum of disease’ – which, in practice, mainly means the distribution of the times since infection in the population being tested. In exploring study design options, we note the large number of assumptions that are required to fully specify details of conventional ‘power calculations’, suggesting that this is not a clear cut approach to choosing sample sizes. Given data generated by a well-understood biological process, we find that formal criteria driven by automated methods for optimising analysis, such as the least absolute shrinkage and selection operator (LASSO), provides little or no advantage over intuitively chosen diagnostic threshold criteria. A detailed mapping of data fields, linking the DS-TB and DR-TB treatment databases is produced, supporting consistent analysis across both types of TB, and facilitating analysis of some of the rich but complex structures in the DR-TB database. Although a significant number of drug-resistant TB patients do not have a recorded treatment outcome, unfavourable treatment outcomes, such as death, are found to be alarmingly common, and significantly associated with HIV status, history of previous TB treatment, age, and resistance patterns. Better management of TB, a persistent and complex infectious disease, will require substantial additional research to be conducted in the coming years. It is hoped that this thesis will provide a meaningful resource to workers in this field, assisting them with numerous aspects of the search for better characterisation of the spectrum of TB; in particular, ways to access and analyse critical data and optimally design data gathering and analysis.
AFRIKAANSE OPSOMMING: daar steeds ‘n gebrek aan ‘n voldoende metode om pasiënte met kliniese aktiewe TB te identifiseer. Die laasgenoemde het twee belangrike negatiewe gevolge: Dikwels bly pasiënte wat dit benodig sonder behandeling, en dikwels word TB behandeling begin as ‘n voorsorg maatreël, veral onder kinders, ten spyte van die gebrek van bewyse van aktiewe TB. In hierdie tesis word verskillende aspekte van die uitdaging om die spektrum van TB-siekte toe te lig, ondersoek: Vernaamste konsepte in diagnose, en die definisie van prestasiemetings vir diagnostiese toetse is ondersoek; ons beskou die uitdagings om ‘n studie te beplan om ‘n kandidaat diagnostiese platform te evalueer; gesimuleerde data, met ‘n gelyksoortige struktuur as ‘n groot voortdurende studie, is ontleed om opsies vir analise en interpretasie van data voortgebring in die begin stadiums van TB-merker evaluasie, te demonstreer; databasis skemas vir die Suid-Afrikaanse openbare sektor se TB-behandelingsprogram word ondersoek, spesifiek om die verskillende bestuur van medikasie-sensitiewe TB (DS-TB) en medikasieweerstandige TB (DR-TB) in verband te bring; ‘n voorlopige ontleding van DR-TB-behandelingsuitkomste word aangebied. Ons demonstreer nie net dat sensitiwiteit en spesifisiteit alleenlik ‘n beperkte oorsig van die verrigting van ‘n toets voorstel nie, maar ook dat hierdie maatstawwe nie ’n intrinsieke deel van ’n toets is nie, maar dat dit varieer met die voorkoms van siektes. In die algemeen, die kontekstuele ‘spektrum van siekte’ – wat in praktyk hoofsaaklik die verspreiding van tye vanaf infeksie in die beproefde populasie beteken. By die ondersoek van opsies vir studieontwerp, neem ons kennis van die groot aantal aannames wat benodig word om die besonderhede ten volle te spesifiseer vir konvensionele onderskeidingvermoë analises, wat daarop dui dat dit nie ’n duidelike benadering tot die keuse van steekproefgroottes is nie. Gegewe data wat gegenereer word deur ’n goed verstaanbare biologiese proses, vind ons dat formele kriteria gedryf deur outomatiese metodes vir die optimalisering van analise, soos die minste absolute krimp- en seleksieoperateur (LASSO), min of geen voordeel bied bo intuïtief gekose diagnostiese drempelkriteria. Die DS-TB en DR-TB behandelingdatabasis word gekoppel deur ‘n gedetailleerde belyning van die datavelde, wat die konsekwente analise van albei tipe TB ondersteun en die analise van sommige van die ryk maar ingewikkelde strukture in die DR-TB databasis fasiliteer. Alhoewel ‘n beduidende aantal pasiënte met medikasie-weerstandige TB geen aangetekende uitkoms van behandeling het nie, word daar gevind dat ongunstige behandelingsuitkomste, soos die dood, onrusbarend algemeen voorkom en dat dit sterk verband hou met MIV-status, geskiedenis van vorige TB-behandeling, ouderdom en patrone in weerstandigheid. In komende jare sal aansienlike bykomende navorsing gedoen moet word om TB, ‘n volhardende en ingewikkelde oordraagbare siekte, beter te bestuur. Daar word gehoop dat hierdie tesis ’n sinvolle hulpbron aan werkers op hierdie gebied sal bied en hulle sal help met talle aspekte van die soeke na beter karakterisering van die spektrum van TB; veral maniere om kritiese data te bekom en te ontleed, en om data-insameling en -ontleding optimaal te ontwerp.
Description
Thesis (MSc)--Stellenbosch University, 2020.
Keywords
Tuberculosis -- Diagnosis -- Mathematical models, Biochemical markers -- Data processing, Tuberculosis -- Spectra -- Management, UCTD
Citation